PDDL

Provides NLP-based PDDL planning, problem generation, batch processing, and execution reporting.

python

3

GitHub Stars

python

Language

5 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "nbnbtm-pddl-mcp-server": {
      "command": "python",
      "args": [
        "d:/PDDL/pddl-mcp/server.py"
      ],
      "env": {
        "FAST_DOWNWARD_PATH": "/path/to/fast-downward.py"
      }
    }
  }
}

This PDDL MCP Server converts natural language planning tasks into PDDL planning problems, executes plans, and supports batch processing. It integrates with Fast Downward and provides type-safe configuration, enabling you to generate plans for single or multiple tasks with detailed execution reports.

How to use

You interact with the server through an MCP client to submit natural language planning tasks and retrieve generated plans. You can create single robot plans, coordinate multiple robots, and run batch task processing to produce comprehensive reports. The server produces PDDL problem files, plan outputs, and explanations to help you understand each step of the planning process.

How to install

Prerequisites: ensure you have Python 3.8+ and a working Internet connection to install dependencies.

Step 1: Install dependencies.

pip install -r requirements.txt

Step 2: Configure Fast Downward.

git clone https://github.com/aibasel/downward.git
cd downward
./build.py

Step 3: Environment configuration.

# Copy this example to the active configuration
cp .env.example .env

Then edit the file to set the paths used by the server.

## Additional sections

Configuration and runtime details are designed to be explicit and actionable for you. You’ll set environment variables that point to the Fast Downward launcher and the PDDL domain templates, then start the server to accept tasks from MCP clients.

A typical local development setup
- Ensure FAST\_DOWNWARD\_PATH points to the fast-downward.py launcher
- Ensure PDDL\_DOMAIN\_PATH points to your domain templates directory
- Start the server in the project directory so the server can locate templates and output folders

## Troubleshooting

If the server doesn’t start or tasks fail to generate, check the following areas:

- Dependencies are installed (`pip install -r requirements.txt`).
- Python version is compatible (3.8+).
- `.env` is properly configured with FAST\_DOWNWARD\_PATH and PDDL\_DOMAIN\_PATH.
- Output directories have write permissions and enough disk space.
- Fast Downward is built and accessible at the path you provided.

## Available tools

### nlp\_to\_pddl

Converts natural language task descriptions into PDDL planning tasks and templates.

### generate\_pddl\_problem

Creates PDDL problem files from task parameters and domain templates.

### batch\_process\_tasks

Runs multiple planning tasks in batch and generates a consolidated execution report.

### validate\_config

Validates configuration and environment before starting planning tasks.